We calibrate our model to scenario 1 in the absence of segregation and coexistence costs, which then constitutes the model baseline. We use the observed and derived prices and quantities for the European Union in the year 2013. We calibrate to scenario 1 as this is the most general scenario, in which NPBTs are regulated as GM and both mandatory GM and voluntary non- GM labeling schemes are in place. The calibration to the most general scenario makes it possible to use the calibrated parameters later in simulating the other scenarios.
In scenario 1, NPBT-derived crops are categorized GM and conventionally produced crops are considered non-GM. But since up to now, all rapeseed in Europe is conventional (and therefore non-GM), we assume for the calibration that the observed prices are non-GM commodity prices ( N R P , N O P , and N M
P ) but that the observed quantities are GM and non-GM quantities. From this assumption, we calculate the equilibrium GM-categorized NPBT prices
G R P , G O P , and N O P .
We assume that the price for rapeseed derived by NPBTs is lower than the conventional rapeseed price, because NPBT crops are produced at lower marginal costs. Estimates of the variable cost differences, for example, for GM and non-GM canola in Canada show mixed results; benefits, such as easier weed control and better time management, are often difficult to quantify (Qaim, 2009; Smyth et al., 2011a). Yield increases and cost reductions through reduced expenditures on herbicides, fuel, and labor have been reported for herbicide-resistant canola in Canada, the United States, and Australia to be higher for the more recent years as compared to the early years after the introduction (Brookes and Barfoot, 2016). We assume a 10-percent cost advantage for GM rapeseed, which represents an average estimate for GM canola for the years 2004 to 2014 as reported by Brookes and Barfoot (2016). The cost advantage implies PRG PRN /1.10 and is assumed to be a result of differences in production costs for competitive farmers, whereas coexistence and segregation costs are assumed to be zero in the calibration.
We assume an equal percentage price advantage for GM oil and meal as compared to their non-GM counterparts. The estimated price advantage must be such, that the crushing costs of
84 GM and non-GM crops are equal. Denoting the relative price premium by x, GM oil and meal prices in the absence of segregation costs (i.e., sO sM ) satisfy 0 G N / (1 )
O O
P P x and
/ (1 )
G N
M M
P P x , respectively. To meet the non-GM zero-profit condition of rapeseed processors in equation (10), the premium is found by rewriting the GM zero-profit condition in equation (9) intoPRG
O OPN MPMN
/ 1
. Using the observed prices x
cR NO
P and N M
P and recalling thatPRG PRN /1.10, we obtain, x 8.8 percent. We assume that the price a processor pays for rapeseed equals the price a farmer receives.
Table 2. Values of Technical Coefficients, Prices, Crushing Costs, and Number of Farmers for the Model Calibration
Description Symbol Value Source/explanation
Oil yield from crushing one metric ton of rapeseed (metric tons)
O
0.38a Ferchau (2000) and FEDIOL (2013) Meal yield from crushing one metric ton of
rapeseed (metric tons)
M
0.62a 1O
Liters of biodiesel from a metric ton of rapeseed oil
B
1,098.08 CARD (2016) Price of GM rapeseed (€/metric ton) G
R
P 386.59 G N /1.10
R R
P P Price of non-GM rapeseed (€/metric ton) N
R
P 425.25 Average price for 2013, UFOP (2013) Price of GM oil (€/metric ton) G
O
P 755.46 G N /1.088
O O
P P Price of non-GM oil (€/metric ton) N
O
P 822.17 Average price for 2013, UFOP (2013) Price of GM meal (€/metric ton) G
M
P 243.12 G N /1.088
M M
P P Price of non-GM meal (€/metric ton) N
M
P 264.58 Average price for 2013, UFOP (2013)
Crushing cost (€/metric ton) cR 51.20 cROPON MPMN PRN
Total number of farmers Z 100.00 Assumed
Number of GM farmers k 67.80 Calculated
Note: a The amount of oil and meal from crushing rapeseed can vary, depending on the type of rapeseed crushing/pressing.
Table 2 summarizes the values of technical coefficients, prices, crushing costs, and the number of GM farmers used to calibrate the model to scenario 1. The number of GM farmers, k, can be thought of as a percentage of the total number of rapeseed farmers, Z, when Z 100.
85 The number of GM farmers is endogenously determined in the calibration (Appendix 5.10.1). Changing the total number of farmers would affect k but not the share of GM farmers,k Z/ .
Rapeseed contains about 43 to 46 percent oil. However, not all oil is extracted during crushing. The extracted oil amount varies between 30 and 43 percent, depending on the type of crushing and pressing of the rapeseed (Ferchau, 2000; Grau et al., 2010). We set the technical oil and meal coefficients to O 0.38 and M 1 0.38 0.62 , respectively. Using the observed non-GM prices as well as the technical oil and meal coefficients, we derive the crushing costs from the zero-profit condition in equation (10). These derived crushing costs are 51.20 euros per metric ton, which is in line with estimates by Ferchau (2000).
The total rapeseed net-supply in 2013 was 25.09 million metric tons (European Commission, 2014). After rapeseed crushing, 2.80 million tons of oil were demanded as food for human consumption. The oil used for biodiesel consumption is calculated by multiplying the share of rapeseed oil in total biodiesel feedstock of 55.67 percent (USDA FAS, 2015b) by the total amount of vegetable oil, 8.51 million tons (FEDIOL, 2013) that was used as feedstock for biodiesel. This calculation yields a biodiesel quantity of 5,202 million liters derived from 4.74 million tons of rapeseed oil. To meet the total rapeseed net-supply we categorize the remaining rapeseed oil of 1.99 million tons as demand for industrial use.
Table 3. Supply and Demand Quantities for Equilibrium Model Calibration
Description Symbol Value Source/explanation a
Total supply of GM rapeseed (metric tons) G R
kS 17.72 Calculated b Total supply of non-GM rapeseed (metric tons)
NR
Zk S 7.37 Calculated b Demand for oil for human consumption
(metric tons)
H O
D 2.80 FEDIOL (2013)
Demand for oil for industrial consumption (metric tons)
I O
D 1.99 Calculated Oil for biodiesel demand (metric tons) B/ B 4.74 USDA FAS
(2015b) and FEDIOL (2013) Demand for meal GM (metric tons) G
M
D 10.98 Calculated Demand for meal non-GM (metric tons) N
M
D 4.57 Calculated
Note: a See text for further explanation on the calculations, b The sum of calculated GM and non-GM rapeseed supply equals 2013 rapeseed supply (USDA FAS, 2015b).
By applying the technical coefficients, crushing and pressing of the total rapeseed net- supply yields 15.55 million tons of meal of which 10.98 million tons are calculated (using the model equations) to be GM and the remainder, 4.57 million tons, is non-GM meal. Given the
86 different demands, the division of rapeseed into GM and non-GM can be derived from the baseline (scenario 1) equation system to be 17.52 and 7.57 million tons, respectively. Table 3 summarizes the supply and demand quantities used in the calibration.
Supply and demand elasticities are taken from the FAPRI elasticity database.43 We use constant price elasticity supply curves for GM and non-GM rapeseed. For a sensitivity analysis, we take these elasticities as the mean values of a beta distribution (Davis, 2008) from which random values are drawn in 10,000 simulations.Table 4 shows the supply and demand elasticity parameters as well as the mean, minimum, and maximum value of the beta distribution. One of the restrictions in our sensitivity analysis is that the own-price elasticity of GM rapeseed supply must be greater than the own-price elasticity of non-GM rapeseed supply. This requirement reflects the effect of the NPBT in lowering the marginal production costs. Furthermore, own- and cross-price elasticities for meal demand are chosen to satisfy the restrictions imposed on the parameters of the underlying utility function.
Table 4. Parameters and Baseline Elasticity Values for Model Calibration
Description Parameter Mean Min Max
Own-price elasticity of GM rapeseed supply G R
0.35b 0.10 0.80
Own-price elasticity of non-GM rapeseed supply N R
0.30a 0.10 0.80
Own-price elasticity of GM oil demand for industrial use
I O
-0.38a -1.00 -0.10
Own-price elasticity non-GM rapeseed oil demand for human consumption
H O
-0.25a -1.00 -0.10 Own-price elasticity of GM meal demand G
M
-4.50b -5.00 -0.80 Own-price elasticity of non-GM meal demand N
M
-4.50b -5.00 -0.80 Cross-price elasticity of demand crossNG
M
0.35b 0.01 1.00
Source: a FAPRI (2013), b assumed to satisfy the conditions of the quasi-linear utility function for vertical product differentiation.